Gruenerator MCP Server
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Gruenerator MCP ServerSearch for renewable energy in the German Green Party program"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Gruenerator MCP Server
A Model Context Protocol (MCP) server that gives AI assistants direct access to Green Party political programs from Germany and Austria.
Table of Contents
Related MCP server: Glean MCP Server
Demo
Coming soon: Demo of search functionality in action
Features
Hybrid Search - Combines vector and text search with RRF fusion
German Optimization - Umlaut handling (ä→ae, ö→oe, etc.) and query variants
Quality Scoring - Results weighted by document quality
Semantic Caching - Fast responses for repeated queries
Metadata Filtering - Filter by document type, section, category, etc.
Filter Discovery - Query available filter values before searching
MCP Resources - Direct access to collection information and AI guidance
Available Document Collections
Collection | Description | Filters |
| Die Grünen Austria: EU Election, Basic Program, National Council |
|
| Bündnis 90/Die Grünen: Basic Program, EU Election, Government Program |
|
| Green Parliamentary Group: Positions and expert content |
|
| gruene.de website: Positions, topics, and news |
|
| gruene.at website: Positions, topics, and news |
|
| KommunalWiki: Municipal politics knowledge (Heinrich Böll Foundation) |
|
| Heinrich-Böll-Stiftung: Analyses, dossiers, and atlases |
|
Prerequisites
Before you begin, ensure you have:
Node.js >= 18.0.0
Qdrant vector database instance (cloud or self-hosted)
Mistral API key for embedding generation (get one here)
Installation
With Docker (Recommended)
# Build image
docker build -t gruenerator-mcp .
# Run container
docker run -d \
--name gruenerator-mcp \
-p 3000:3000 \
-e QDRANT_URL=https://your-qdrant.com \
-e QDRANT_API_KEY=your-api-key \
-e MISTRAL_API_KEY=your-mistral-key \
gruenerator-mcpLocal Development
# Clone repository
git clone https://github.com/Movm/Gruenerator-MCP.git
cd Gruenerator-MCP
# Install dependencies
npm install
# Set environment variables
cp .env.example .env
# Edit .env with your credentials
# Start server
npm start
# Or with auto-reload
npm run devWith Coolify
Create a new project
Connect Git repository:
https://github.com/Movm/Gruenerator-MCPSet environment variables (see below)
Deploy
Configuration
Environment Variables
Variable | Description | Required |
| URL to Qdrant instance | Yes |
| API key for Qdrant | Yes |
| API key for Mistral embeddings | Yes |
| Server port | No (default: 3000) |
| Public URL for config generation | No |
| Log level: DEBUG, INFO, WARN, ERROR | No (default: INFO) |
MCP Client Setup
Cursor / Claude Desktop
Add to your MCP configuration:
{
"mcpServers": {
"gruenerator": {
"url": "http://localhost:3000/mcp"
}
}
}After configuration, you can ask questions like:
"Search the Austrian Green programs for climate policy"
"What does the German Green basic program say about energy transition?"
"Find municipal politics guidance on waste management in KommunalWiki"
"What are the Bundestagsfraktion's positions on renewable energy?"
"Search Böll-Stiftung for analyses on feminism and gender equality"
API Reference
Endpoints
Endpoint | Method | Description |
| POST | MCP communication |
| GET | SSE stream |
| DELETE | End session |
| GET | Health check with cache and request statistics |
| GET | Detailed server metrics |
| GET | Auto-discovery metadata |
| GET | Generate client configuration |
| GET | Server information |
Health Check Response
{
"status": "ok",
"service": "gruenerator-mcp",
"version": "1.0.0",
"collections": ["oesterreich", "deutschland", "bundestagsfraktion", "gruene-de", "gruene-at", "kommunalwiki", "boell-stiftung"],
"uptime": { "ms": 3600000, "hours": 1.0 },
"cache": {
"embeddingHitRate": "65%",
"searchHitRate": "42%"
},
"requests": { "total": 150, "searches": 120, "errors": 0 },
"performance": { "avgResponseTimeMs": 250, "cacheHitRate": "65%" }
}Metrics Response
{
"server": { "name": "gruenerator-mcp", "version": "1.0.0" },
"uptime": { "hours": 1.0 },
"requests": { "total": 150, "searches": 120 },
"breakdown": {
"byCollection": { "deutschland": 40, "oesterreich": 30, "kommunalwiki": 25, "bundestagsfraktion": 20, "gruene-de": 15, "gruene-at": 20 },
"bySearchMode": { "hybrid": 100, "vector": 15, "text": 5 }
},
"cache": {
"embeddings": { "entries": 50, "hitRate": "65%" },
"search": { "entries": 30, "hitRate": "42%" }
},
"memory": { "heapUsedMB": 45, "rssMB": 120 }
}MCP Tools & Resources
Tools
gruenerator_search
Searches party programs and content with hybrid, vector, or text search.
Parameters:
Parameter | Type | Description | Default |
| string | Search term or question | required |
| string | Collection to search (see table above) | required |
| string |
|
|
| number | Maximum number of results (1-20) | 5 |
| object | Metadata filters (see below) | optional |
| boolean | Use cache | true |
Available Filters by Collection:
Filter | Collections | Description |
| oesterreich, deutschland | Document type (grundsatzprogramm, wahlprogramm, eu-wahlprogramm) |
| oesterreich, deutschland | Exact document title |
| bundestagsfraktion, gruene-de, gruene-at | Content section (Positionen, Themen, etc.) |
| kommunalwiki | Article type (literatur, praxishilfe, faq, etc.) |
| kommunalwiki | Thematic category (Haushalt, Umwelt, etc.) |
| boell-stiftung | Content type (artikel, dossier, atlas, schwerpunkt) |
| boell-stiftung | Topic (klima, feminismus, migration, digitalisierung, etc.) |
| boell-stiftung | Geographic region (afrika, europa, asien, lateinamerika, etc.) |
Example:
{
"query": "climate protection and renewable energy",
"collection": "kommunalwiki",
"searchMode": "hybrid",
"limit": 5,
"filters": { "article_type": "praxishilfe", "category": "Umwelt" }
}gruenerator_get_filters
Discovers available filter values for a collection. Use this before filtering to know valid values.
Parameters:
Parameter | Type | Description | Default |
| string | Collection to get filters for | required |
Example Response:
{
"collection": "KommunalWiki",
"collectionId": "kommunalwiki",
"description": "Municipal politics knowledge (Heinrich Böll Foundation)",
"filters": {
"article_type": {
"label": "Article Type",
"type": "keyword",
"values": ["literatur", "praxishilfe", "faq", "sachgebiet"],
"count": 4
},
"category": {
"label": "Category",
"type": "keyword",
"values": ["Haushalt", "Umwelt", "Verkehr", "..."],
"count": 25
}
}
}gruenerator_cache_stats
Shows cache statistics for embeddings and search results.
Parameters: None
Example Response:
{
"embeddings": { "entries": 50, "hits": 120, "misses": 30, "hitRate": "80%" },
"search": { "entries": 30, "hits": 80, "misses": 40, "hitRate": "67%" }
}get_client_config
Generates ready-to-use MCP configurations for various clients.
Parameters:
client(string, required) -claude,cursor, orvscode
Resources
The server provides the following resources via MCP protocol:
URI | Description |
| Read first! Usage instructions for AI assistants |
| Server information and capabilities |
| List of all available collections |
| Details of a specific collection (oesterreich, deutschland, bundestagsfraktion, gruene-de, gruene-at, kommunalwiki, boell-stiftung) |
Search Modes
Hybrid (Default)
Combines vector and text search with Reciprocal Rank Fusion (RRF). Best results for most queries.
Vector
Pure semantic search based on embeddings. Good for conceptual questions.
Text
Classic text search with German optimizations. Good for exact terms or names.
Contributing
Contributions are welcome! Here's how you can help:
Report bugs - Open an issue using the bug report template
Request features - Open an issue using the feature request template
Submit PRs - Fork the repo, create a branch, and submit a pull request
Please ensure your PR:
Passes CI checks
Follows existing code style
Includes appropriate documentation updates
Acknowledgments
Model Context Protocol - The protocol powering this server
Qdrant - Vector database for semantic search
Mistral AI - Embedding generation
Die Grünen Österreich & Bündnis 90/Die Grünen - Source documents
Grüne Bundestagsfraktion - Parliamentary group content
Heinrich-Böll-Stiftung - Political analyses, dossiers, and atlases
Heinrich-Böll-Stiftung KommunalWiki - Municipal politics knowledge base
License
MIT License - see LICENSE
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Movm/Gruenerator-MCP'
If you have feedback or need assistance with the MCP directory API, please join our Discord server